George K.I. Mann, Prodyut K. Roy, and Bipul C. Hawlader
Model predictive control, adaptive control, multivariable control, adaptive fuzzy model, dynamic matrix control
This paper presents a novel approach to design an adaptive fuzzy model-based predictive control (MPC) algorithm for controlling temperatures and its gradient of a multivariable soil-heating process system. The model of the system uses Takagi–Sugeno (TS) type fuzzy inference structure. The TS rules are described in the parametric form to realize recursive least square (RLS) method for online identification of the TS model. The control objective is to track a desired temperature profile at three different locations in three different zones in a soil cell. Three heat sources are located at the outer surface of the soil cell. In each sampling instance, the system identifies the fuzzy rules and they are recursively adapted for handling the time-variant behaviour of the process. For simulations, the soil-heating system is modelled using a finite element (FE) program, ABAQUS. The dynamic control program is linked to the FE system using a user-defined subroutine. The proposed fuzzy adaptive MPC scheme is compared against the non-adaptive fuzzy MPC scheme. Further, the system is also compared against the classical MPC scheme to confirm the superiority of the proposed algorithm.
Important Links:
Go Back
